Search

In this paper, the privacy problem of a parallel distributed detection system vulnerable to an eavesdropper is proposed and studied in the Neyman-Pearson formulation. The privacy leakage is evaluated by a metric related to the Neyman-Pearson criterion. We will show that it is sufficient to consider a deterministic likelihood-ratio test for the optimal detection strategy at the eavesdropped sensor. This fundamental insight helps to simplify the problem to find the optimal privacy-constrained distributed detection system design. The trade-off between the detection performance and privacy leakage is illustrated in a numerical example., QC 20150122. QC 20160314

In this paper, the privacy problem of a parallel distributed detection system vulnerable to an eavesdropper is proposed and studied in the Neyman-Pearson formulation. The privacy leakage is evaluated by a metric related to the Neyman-Pearson criterion. We will show that it is sufficient to consider a deterministic likelihood-ratio test for the optimal detection strategy at the eavesdropped sensor. This fundamental insight helps to simplify the problem to find the optimal privacy-constrained distributed detection system design. The trade-off between the detection performance and privacy leakage is illustrated in a numerical example., QC 20150122. QC 20160314

This paper presents for the first time the testbed implementation of an iterative interference alignment and power control algorithm for downlink transmission in a multiple-input multiple-output (MIMO) cellular network. The network is composed of three cells where within each cell one base station (BS) communicates with one mobile station (MS). Each terminal is equipped with two antennas. All the BSs transmit at the same time and the same frequency band. Transmitter beamforming vectors and receiver filtering vectors are computed according to the interference alignment concept, and power control is performed to guarantee successful communication of each BS-MS pair at a desired fixed rate. The indoor measurements performed on an universal software radio peripheral (USRP) based test-bed, show that the power can be reduced by at least 4 dB, 90% of the time, while at the same time reducing the bit-error-rate (BER)., QC 20150227

The computation of mutual informations of large scale systems with finite input alphabet and Gaussian noise has often prohibitive complexities. In this paper, we propose a novel approach exploiting the sphere decoding concept to bound and approximate such mutual information term with reduced complexity and good accuracy. Using Monte-Carlo simulations, the method is numerically demonstrated for the computation of the mutual information of a frequency- and time-selective channel with QAM modulation., QC 20150604

This paper considers the capacity limit of an uplink wideband CDMA system. In order to make the studied results useful for the performance assessment of real cellular networks, various realistic assumptions are included in the problem. An equivalent discrete-time channel model is derived based on sufcient statistic for optimal decoding of transmit messages. The capacity regions are then characterized considering nite constellation and Gaussian input assumptions. For further insight, an analysis on the asymptotic capacity is considered, in which the conditions to simultaneously achieve the individual capacities are derived., QC 20150608

In this paper, the privacy problem of a parallel distributed detection system vulnerable to an eavesdropper is proposed and studied in the Neyman-Pearson formulation. The privacy leakage is evaluated by a metric related to the Neyman-Pearson criterion. We will show that it is sufficient to consider a deterministic likelihood-ratio test for the optimal detection strategy at the eavesdropped sensor. This fundamental insight helps to simplify the problem to find the optimal privacy-constrained distributed detection system design. The trade-off between the detection performance and privacy leakage is illustrated in a numerical example., QC 20150122. QC 20160314

In this paper, the privacy problem of a parallel distributed detection system vulnerable to an eavesdropper is proposed and studied in the Neyman-Pearson formulation. The privacy leakage is evaluated by a metric related to the Neyman-Pearson criterion. We will show that it is sufficient to consider a deterministic likelihood-ratio test for the optimal detection strategy at the eavesdropped sensor. This fundamental insight helps to simplify the problem to find the optimal privacy-constrained distributed detection system design. The trade-off between the detection performance and privacy leakage is illustrated in a numerical example., QC 20150122. QC 20160314

In this paper, the privacy problem of a parallel distributed detection system vulnerable to an eavesdropper is proposed and studied in the Neyman-Pearson formulation. The privacy leakage is evaluated by a metric related to the Neyman-Pearson criterion. We will show that it is sufficient to consider a deterministic likelihood-ratio test for the optimal detection strategy at the eavesdropped sensor. This fundamental insight helps to simplify the problem to find the optimal privacy-constrained distributed detection system design. The trade-off between the detection performance and privacy leakage is illustrated in a numerical example., QC 20150122. QC 20160314

Suppose a linear model Y = Hx+n, where inputs x, n are independent Gaussian mixtures. The problem is to design the transfer matrix so as to minimize the mean square error (MSE) when estimating x from . This problem has important applications, but faces at least three hurdles. Firstly, even for a fixed H, the minimum MSE (MMSE) has no analytical form. Secondly, theMMSE is generally not convex in . Thirdly, derivatives of the MMSEw.r.t. are hard to obtain. This paper casts theproblemas a stochastic program and invokes gradient methods. The study is motivated by two applications in signal processing. One concerns the choice of error-reducing precoders; the other deals with selection of pilot matrices for channel estimation. In either setting, our numerical results indicate improved estimation accuracy-markedly better than those obtained by optimal design based on standard linear estimators. Some implications of the non-convexities of the MMSE are noteworthy, yet, to our knowledge, not well known. For example, there are cases in which more pilot power is detrimental for channel estimation. This paper explains why., QC 20131029

In sparse recovery from measurement data a common approach is to use greedy pursuit reconstruction algorithms. Most of these algorithms have a correlation filter for detecting active components in the sparse data. In this paper, we show how modifications can be made for the greedy pursuit algorithms so that they use beamformers insteadof the standard correlation filter. Using these beamformers, improved performance in the algorithms is obtained. In particular, we discuss beamformers for the average and worst case scenario and give methods for constructing them., QC 20131022

The Capon method is a powerful nonparametric approach in array processing based on the sample covariance matrix. For small sample sets, however, its performance is degraded. In this paper we formulate a regularized covariance matching framework based on the nuclear norm for enhancing the Capon method. An approximate iterative solution is developed and tested using simulated data. Appropriate regularization parameter values are also inferred from the data, drawing upon the cross-validation approach. The results show significantly improved spatial spectral and signal waveform estimates., QC 20140127

In sparse recovery from measurement data a common approach is to use greedy pursuit reconstruction algorithms. Most of these algorithms have a correlation filter for detecting active components in the sparse data. In this paper, we show how modifications can be made for the greedy pursuit algorithms so that they use beamformers insteadof the standard correlation filter. Using these beamformers, improved performance in the algorithms is obtained. In particular, we discuss beamformers for the average and worst case scenario and give methods for constructing them., QC 20131022

In sparse recovery from measurement data a common approach is to use greedy pursuit reconstruction algorithms. Most of these algorithms have a correlation filter for detecting active components in the sparse data. In this paper, we show how modifications can be made for the greedy pursuit algorithms so that they use beamformers insteadof the standard correlation filter. Using these beamformers, improved performance in the algorithms is obtained. In particular, we discuss beamformers for the average and worst case scenario and give methods for constructing them., QC 20131022

In sparse recovery from measurement data a common approach is to use greedy pursuit reconstruction algorithms. Most of these algorithms have a correlation filter for detecting active components in the sparse data. In this paper, we show how modifications can be made for the greedy pursuit algorithms so that they use beamformers insteadof the standard correlation filter. Using these beamformers, improved performance in the algorithms is obtained. In particular, we discuss beamformers for the average and worst case scenario and give methods for constructing them., QC 20131022

In sparse recovery from measurement data a common approach is to use greedy pursuit reconstruction algorithms. Most of these algorithms have a correlation filter for detecting active components in the sparse data. In this paper, we show how modifications can be made for the greedy pursuit algorithms so that they use beamformers insteadof the standard correlation filter. Using these beamformers, improved performance in the algorithms is obtained. In particular, we discuss beamformers for the average and worst case scenario and give methods for constructing them., QC 20131022

In this paper, a method for determining ground plane features in a sequence of images captured by a mobile camera is presented. The hardware of the mobile system consists of a monocular camera that is mounted on an inertial measurement unit (IMU). An image processing procedure is proposed, first to extract image features and match them across consecutive image frames, and second to detect the ground plane features using a two-step algorithm. In the first step, the planar homography of the ground plane is constructed using an IMU-camera motion estimation approach. The obtained homography constraints are used to detect the most likely ground features in the sequence of images. To reject the remaining outliers, as the second step, a new plane normal vector computation approach is proposed. To obtain the normal vector of the ground plane, only three pairs of corresponding features are used for a general camera transformation. The normal-based computation approach generalizes the existing methods that are developed for specific camera transformations. Experimental results on real data validate the reliability of the proposed method., QC 20121107

In this paper, a method for determining ground plane features in a sequence of images captured by a mobile camera is presented. The hardware of the mobile system consists of a monocular camera that is mounted on an inertial measurement unit (IMU). An image processing procedure is proposed, first to extract image features and match them across consecutive image frames, and second to detect the ground plane features using a two-step algorithm. In the first step, the planar homography of the ground plane is constructed using an IMU-camera motion estimation approach. The obtained homography constraints are used to detect the most likely ground features in the sequence of images. To reject the remaining outliers, as the second step, a new plane normal vector computation approach is proposed. To obtain the normal vector of the ground plane, only three pairs of corresponding features are used for a general camera transformation. The normal-based computation approach generalizes the existing methods that are developed for specific camera transformations. Experimental results on real data validate the reliability of the proposed method., QC 20121107

In this paper, a method for determining ground plane features in a sequence of images captured by a mobile camera is presented. The hardware of the mobile system consists of a monocular camera that is mounted on an inertial measurement unit (IMU). An image processing procedure is proposed, first to extract image features and match them across consecutive image frames, and second to detect the ground plane features using a two-step algorithm. In the first step, the planar homography of the ground plane is constructed using an IMU-camera motion estimation approach. The obtained homography constraints are used to detect the most likely ground features in the sequence of images. To reject the remaining outliers, as the second step, a new plane normal vector computation approach is proposed. To obtain the normal vector of the ground plane, only three pairs of corresponding features are used for a general camera transformation. The normal-based computation approach generalizes the existing methods that are developed for specific camera transformations. Experimental results on real data validate the reliability of the proposed method., QC 20121107

In this paper, a method for determining ground plane features in a sequence of images captured by a mobile camera is presented. The hardware of the mobile system consists of a monocular camera that is mounted on an inertial measurement unit (IMU). An image processing procedure is proposed, first to extract image features and match them across consecutive image frames, and second to detect the ground plane features using a two-step algorithm. In the first step, the planar homography of the ground plane is constructed using an IMU-camera motion estimation approach. The obtained homography constraints are used to detect the most likely ground features in the sequence of images. To reject the remaining outliers, as the second step, a new plane normal vector computation approach is proposed. To obtain the normal vector of the ground plane, only three pairs of corresponding features are used for a general camera transformation. The normal-based computation approach generalizes the existing methods that are developed for specific camera transformations. Experimental results on real data validate the reliability of the proposed method., QC 20121107

In this paper, a method for determining ground plane features in a sequence of images captured by a mobile camera is presented. The hardware of the mobile system consists of a monocular camera that is mounted on an inertial measurement unit (IMU). An image processing procedure is proposed, first to extract image features and match them across consecutive image frames, and second to detect the ground plane features using a two-step algorithm. In the first step, the planar homography of the ground plane is constructed using an IMU-camera motion estimation approach. The obtained homography constraints are used to detect the most likely ground features in the sequence of images. To reject the remaining outliers, as the second step, a new plane normal vector computation approach is proposed. To obtain the normal vector of the ground plane, only three pairs of corresponding features are used for a general camera transformation. The normal-based computation approach generalizes the existing methods that are developed for specific camera transformations. Experimental results on real data validate the reliability of the proposed method., QC 20121107

This paper presents a new approach to soft demodulation for MIMO channels. The proposed method is an approximation to the exact a posteriori probability-per-bit computer. The main idea is to marginalize the posterior density for the received data exactly over the subset of the transmitted bits that are received with the lowest signal-to-noise-ratio, and then marginalize this density approximately over the remaining bits. Unlike the exact demodulator, whose complexity is huge due to the need for enumerating all possible combinations of transmitted constellation points, the proposed method has very low complexity. Additionally, its complexity is fixed, which makes it suitable for pipelined implementation. Numerical examples illustrate its performance on slow fading 4 x 4 and 6 x 6 complex MIMO channels.

Analog-to-digital converter additive post-correction using look-up-tables is considered. In a practical post-correction system, the correction values must be stored using limited (fixed-point) precision. In this paper the effects of limited precision in the correction values is investigated. A theory for approximating the SINAD after correction using fixed-point values is presented. The theory shows good agreement when compared with simulation results., NQC

—In this paper we give a brief introduction to some issues and results regarding ADC post-correction. The emphasis is on signal process-ing aspects of post-correction, including dynamics, estimation and round-off effects., NQC

This paper focuses on the problem of reconstructing low-rank matrices from underdetermined measurements using alternating optimization strategies. We endeavour to combine an alternating least-squares based estimation strategy with ideas from the alternating direction method of multipliers (ADMM) to recover low-rank matrices with linear parameterized structures, such as Hankel matrices. The use of ADMM helps to improve the estimate in each iteration due to its capability of incorporating information about the direction of estimates achieved in previous iterations. We show that merging these two alternating strategies leads to a better performance and less consumed time than the existing alternating least squares (ALS) strategy. The improved performance is verified via numerical simulations with varying sampling rates and real applications., QC 20160202. QC 20160304

The entropy computation of Gaussian mixture distributions with a large number of components has a prohibitive computational complexity. In this paper, we propose a novel approach exploiting the sphere decoding concept to bound and approximate such entropy terms with reduced complexity and good accuracy. Moreover, we propose an SNR region-based enhancement of the approximation method to reduce the complexity even further. Using Monte-Carlo simulations, the proposed methods are numerically demonstrated for the computation of the mutual information including the entropy term of various channels with finite constellation modulations such as binary and quadratic amplitude modulation (QAM) inputs for communication applications., QC 20150918

Low-rank matrix reconstruction (LRMR) problem considersestimation (or reconstruction) of an underlying low-rank matrixfrom under-sampled linear measurements. A low-rank matrix can be represented using a factorized model. In thisarticle, we derive Bayesian Cramer-Rao bounds for LRMR where a factorized model is used. We first show a general informative bound, and then derive several Bayesian Cramer-Rao bounds for different scenarios. We always considered the low-rank matrix to be reconstructed as a random matrix, but its model hyper-parameters for three cases - deterministic known, deterministic unknown and random. Finally we compare the bounds with existing practical algorithms through numerical simulations., QC 20160810

Low-rank matrix reconstruction (LRMR) problem considersestimation (or reconstruction) of an underlying low-rank matrixfrom under-sampled linear measurements. A low-rank matrix can be represented using a factorized model. In thisarticle, we derive Bayesian Cramer-Rao bounds for LRMR where a factorized model is used. We first show a general informative bound, and then derive several Bayesian Cramer-Rao bounds for different scenarios. We always considered the low-rank matrix to be reconstructed as a random matrix, but its model hyper-parameters for three cases - deterministic known, deterministic unknown and random. Finally we compare the bounds with existing practical algorithms through numerical simulations., QC 20160810

We consider a decentralized hypothesis testing problem in which several peripheral energy harvesting sensors arearranged in parallel. Each sensor makes a noisy observation of a time varying phenomenon, and sends a message about the present hypothesis towards a fusion center at each time instance t. The fusion center, using the aggregate of the received messages during the time instance t, makes a decision about the state of the present hypothesis. We assume that each sensor is an energy harvesting device and is capable of harvesting all the energy it needs to communicate from its environment. Our contribution is to formulate and analyze the decentralized detection problem when the energy harvesting sensors are allowed to form a long term energy usage policy. Our analysis is based on a queuing-theoretic model for the battery. Then, by using numerical simulations, we show how the resulting performance differs from the energy-unconstrained case., QC 20160907

There has been growing interest in millimeter wave (mmWave) multiple-input multiple-output (MIMO) systems, which would likely employ hybrid analog-digital precoding with large-scale analog arrays deployed at wide bandwidths. Primary challenges here are how to efficiently estimate the large-dimensional frequency-selective channels and customize the wideband hybrid analog-digital precoders and combiners. To address these challenges, we propose a low-overhead channel subspace estimation technique for the wideband hybrid analog-digital MIMO precoding systems. We first show that the Gram matrix of the frequency-selective channel can be decomposed into frequency-flat and frequency-selective components. Based on this, the Arnoldi approach, leveraging channel reciprocity and time-reversed echoing, is employed to estimate a frequency-flat approximation of the frequency-selective mmWave channels, which is used to design the analog parts. After the analog precoder and combiner design, the low-dimensional frequency-selective channels are estimated using conventional pilot-based channel sounding. Numerical results show that considerable improvement in data-rate performance is possible., QC 20170110

We consider a decentralized hypothesis testing problem in which several peripheral energy harvesting sensors are arranged in parallel. Each sensor makes a noisy observation of a time varying phenomenon, and sends a message about the present hypothesis towards a fusion center at each time instance t. The fusion center, using the aggregate of the received messages during the time instance t, makes a decision about the state of the present hypothesis. We assume that each sensor is an energy harvesting device and is capable of harvesting all the energy it needs to communicate from its environment. Our contribution is to formulate and analyze the decentralized detection problem when the energy harvesting sensors are allowed to form a long term energy usage policy. Our analysis is based on a queuing-theoretic model for the battery. Then, by using numerical simulations, we show how the resulting performance differs from the energy-unconstrained case., QC 20170215

Low-rank matrix reconstruction (LRMR) problem considersestimation (or reconstruction) of an underlying low-rank matrixfrom under-sampled linear measurements. A low-rank matrix can be represented using a factorized model. In thisarticle, we derive Bayesian Cramer-Rao bounds for LRMR where a factorized model is used. We first show a general informative bound, and then derive several Bayesian Cramer-Rao bounds for different scenarios. We always considered the low-rank matrix to be reconstructed as a random matrix, but its model hyper-parameters for three cases - deterministic known, deterministic unknown and random. Finally we compare the bounds with existing practical algorithms through numerical simulations., QC 20160810

Low-rank matrix reconstruction (LRMR) problem considersestimation (or reconstruction) of an underlying low-rank matrixfrom under-sampled linear measurements. A low-rank matrix can be represented using a factorized model. In thisarticle, we derive Bayesian Cramer-Rao bounds for LRMR where a factorized model is used. We first show a general informative bound, and then derive several Bayesian Cramer-Rao bounds for different scenarios. We always considered the low-rank matrix to be reconstructed as a random matrix, but its model hyper-parameters for three cases - deterministic known, deterministic unknown and random. Finally we compare the bounds with existing practical algorithms through numerical simulations., QC 20160810

Low-rank matrix reconstruction (LRMR) problem considersestimation (or reconstruction) of an underlying low-rank matrixfrom under-sampled linear measurements. A low-rank matrix can be represented using a factorized model. In thisarticle, we derive Bayesian Cramer-Rao bounds for LRMR where a factorized model is used. We first show a general informative bound, and then derive several Bayesian Cramer-Rao bounds for different scenarios. We always considered the low-rank matrix to be reconstructed as a random matrix, but its model hyper-parameters for three cases - deterministic known, deterministic unknown and random. Finally we compare the bounds with existing practical algorithms through numerical simulations., QC 20160810

We consider a decentralized hypothesis testing problem in which several peripheral energy harvesting sensors arearranged in parallel. Each sensor makes a noisy observation of a time varying phenomenon, and sends a message about the present hypothesis towards a fusion center at each time instance t. The fusion center, using the aggregate of the received messages during the time instance t, makes a decision about the state of the present hypothesis. We assume that each sensor is an energy harvesting device and is capable of harvesting all the energy it needs to communicate from its environment. Our contribution is to formulate and analyze the decentralized detection problem when the energy harvesting sensors are allowed to form a long term energy usage policy. Our analysis is based on a queuing-theoretic model for the battery. Then, by using numerical simulations, we show how the resulting performance differs from the energy-unconstrained case., QC 20160907

We consider a decentralized hypothesis testing problem in which several peripheral energy harvesting sensors arearranged in parallel. Each sensor makes a noisy observation of a time varying phenomenon, and sends a message about the present hypothesis towards a fusion center at each time instance t. The fusion center, using the aggregate of the received messages during the time instance t, makes a decision about the state of the present hypothesis. We assume that each sensor is an energy harvesting device and is capable of harvesting all the energy it needs to communicate from its environment. Our contribution is to formulate and analyze the decentralized detection problem when the energy harvesting sensors are allowed to form a long term energy usage policy. Our analysis is based on a queuing-theoretic model for the battery. Then, by using numerical simulations, we show how the resulting performance differs from the energy-unconstrained case., QC 20160907

We consider a decentralized hypothesis testing problem in which several peripheral energy harvesting sensors arearranged in parallel. Each sensor makes a noisy observation of a time varying phenomenon, and sends a message about the present hypothesis towards a fusion center at each time instance t. The fusion center, using the aggregate of the received messages during the time instance t, makes a decision about the state of the present hypothesis. We assume that each sensor is an energy harvesting device and is capable of harvesting all the energy it needs to communicate from its environment. Our contribution is to formulate and analyze the decentralized detection problem when the energy harvesting sensors are allowed to form a long term energy usage policy. Our analysis is based on a queuing-theoretic model for the battery. Then, by using numerical simulations, we show how the resulting performance differs from the energy-unconstrained case., QC 20160907

We consider a decentralized hypothesis testing problem in which several peripheral energy harvesting sensors arearranged in parallel. Each sensor makes a noisy observation of a time varying phenomenon, and sends a message about the present hypothesis towards a fusion center at each time instance t. The fusion center, using the aggregate of the received messages during the time instance t, makes a decision about the state of the present hypothesis. We assume that each sensor is an energy harvesting device and is capable of harvesting all the energy it needs to communicate from its environment. Our contribution is to formulate and analyze the decentralized detection problem when the energy harvesting sensors are allowed to form a long term energy usage policy. Our analysis is based on a queuing-theoretic model for the battery. Then, by using numerical simulations, we show how the resulting performance differs from the energy-unconstrained case., QC 20160907

In this work we shed light on the problem of precoding and user selection in MIMO networks. We formulate the problem using the framework of stable matching, whereby a set of users wish to be matched to a set of serving base stations, such as to maximize the sum-rate performance of the system. Though the problem is NP-hard, we propose a suboptimal heuristic that tackles the problem in a distributed fashion: we apply a many-to-one stable matching algorithm to generate a sequence of matchings, and the Weighted MMSE algorithm to perform the precoding. We benchmark our algorithm againt the recently proposed Weighted MMSE with User Assignment algorithm [1]., QC 20160613

We develop a greedy algorithm for the basis-pursuit problem. Thealgorithm is empirically found to provide the same solution as convex optimization based solvers. The method uses only a subset ofthe optimization variables in each iteration and iterates until an optimality condition is satisfied. In simulations, the algorithm converges faster than standard methods when the number of measurements is small and the number of variables large., Presented at the 40th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2015 in Brisbane, Australia. QC 20150415

We develop a greedy algorithm for the basis-pursuit problem. Thealgorithm is empirically found to provide the same solution as convex optimization based solvers. The method uses only a subset ofthe optimization variables in each iteration and iterates until an optimality condition is satisfied. In simulations, the algorithm converges faster than standard methods when the number of measurements is small and the number of variables large., Presented at the 40th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2015 in Brisbane, Australia. QC 20150415

In this work, we address the problem of channel estimation and precoding / combining for the so-called hybrid millimeter wave (mmWave) MIMO architecture. Our proposed channel estimation scheme exploits channel reciprocity in TDD MIMO systems, by using echoing, thereby allowing us to implement Krylov subspace methods in a fully distributed way. The latter results in estimating the right (resp. left) singular subspace of the channel at the transmitter (resp. receiver). Moreover, we also tackle the problem of subspace decomposition whereby the estimated right (resp. left) singular subspaces are approximated by a cascade of analog and digital precoder (resp. combiner), using an iterative method. Finally we compare our scheme with an equivalent fully digital case and conclude that a relatively similar performance can be achieved, however, with a drastically reduced number of RF chains - 4 ~ 8 times less (i.e., massive savings in cost and power consumption)., QC 20151217

We consider a MEMO interference channel in which the transmitters and receivers operate in frequency-division duplex mode. In this setting, interference management through coordinated transceiver design necessitates channel state information at the transmitters (CSI-T). The acquisition of CSI T is done through feedback from the receivers, which entitles a loss in degrees of freedom, due to training and feedback. This loss increases with the amount of CSI-T. In this work, after formulating an overhead model for CSI acquisition at the transmitters, we propose a distributed mechanism to find for each transmitter a subset of the complete CSI, which is used to perform interference management. The mechanism is based on many-to-many stable matching. We prove the existence of a stable matching and exploit an algorithm to reach it. Simulation results show performance improvement compared to full and minimal CSI-T., QC 20160715

We develop a greedy algorithm for the basis-pursuit problem. Thealgorithm is empirically found to provide the same solution as convex optimization based solvers. The method uses only a subset ofthe optimization variables in each iteration and iterates until an optimality condition is satisfied. In simulations, the algorithm converges faster than standard methods when the number of measurements is small and the number of variables large., Presented at the 40th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2015 in Brisbane, Australia. QC 20150415

We develop a greedy algorithm for the basis-pursuit problem. Thealgorithm is empirically found to provide the same solution as convex optimization based solvers. The method uses only a subset ofthe optimization variables in each iteration and iterates until an optimality condition is satisfied. In simulations, the algorithm converges faster than standard methods when the number of measurements is small and the number of variables large., Presented at the 40th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2015 in Brisbane, Australia. QC 20150415

In this work, we address the problem of channel estimation and precoding / combining for the so-called hybrid millimeter wave (mmWave) MIMO architecture. Our proposed channel estimation scheme exploits channel reciprocity in TDD MIMO systems, by using echoing, thereby allowing us to implement Krylov subspace methods in a fully distributed way. The latter results in estimating the right (resp. left) singular subspace of the channel at the transmitter (resp. receiver). Moreover, we also tackle the problem of subspace decomposition whereby the estimated right (resp. left) singular subspaces are approximated by a cascade of analog and digital precoder (resp. combiner), using an iterative method. Finally we compare our scheme with an equivalent fully digital case and conclude that a relatively similar performance can be achieved, however, with a drastically reduced number of RF chains - 4 ~ 8 times less (i.e., massive savings in cost and power consumption)., QC 20151217

We develop a greedy algorithm for the basis-pursuit problem. Thealgorithm is empirically found to provide the same solution as convex optimization based solvers. The method uses only a subset ofthe optimization variables in each iteration and iterates until an optimality condition is satisfied. In simulations, the algorithm converges faster than standard methods when the number of measurements is small and the number of variables large., Presented at the 40th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2015 in Brisbane, Australia. QC 20150415

In this work, we address the problem of channel estimation and precoding / combining for the so-called hybrid millimeter wave (mmWave) MIMO architecture. Our proposed channel estimation scheme exploits channel reciprocity in TDD MIMO systems, by using echoing, thereby allowing us to implement Krylov subspace methods in a fully distributed way. The latter results in estimating the right (resp. left) singular subspace of the channel at the transmitter (resp. receiver). Moreover, we also tackle the problem of subspace decomposition whereby the estimated right (resp. left) singular subspaces are approximated by a cascade of analog and digital precoder (resp. combiner), using an iterative method. Finally we compare our scheme with an equivalent fully digital case and conclude that a relatively similar performance can be achieved, however, with a drastically reduced number of RF chains - 4 ~ 8 times less (i.e., massive savings in cost and power consumption)., QC 20151217

In this work, we address the problem of channel estimation and precoding / combining for the so-called hybrid millimeter wave (mmWave) MIMO architecture. Our proposed channel estimation scheme exploits channel reciprocity in TDD MIMO systems, by using echoing, thereby allowing us to implement Krylov subspace methods in a fully distributed way. The latter results in estimating the right (resp. left) singular subspace of the channel at the transmitter (resp. receiver). Moreover, we also tackle the problem of subspace decomposition whereby the estimated right (resp. left) singular subspaces are approximated by a cascade of analog and digital precoder (resp. combiner), using an iterative method. Finally we compare our scheme with an equivalent fully digital case and conclude that a relatively similar performance can be achieved, however, with a drastically reduced number of RF chains - 4 ~ 8 times less (i.e., massive savings in cost and power consumption)., QC 20151217

In this work, we address the problem of channel estimation and precoding / combining for the so-called hybrid millimeter wave (mmWave) MIMO architecture. Our proposed channel estimation scheme exploits channel reciprocity in TDD MIMO systems, by using echoing, thereby allowing us to implement Krylov subspace methods in a fully distributed way. The latter results in estimating the right (resp. left) singular subspace of the channel at the transmitter (resp. receiver). Moreover, we also tackle the problem of subspace decomposition whereby the estimated right (resp. left) singular subspaces are approximated by a cascade of analog and digital precoder (resp. combiner), using an iterative method. Finally we compare our scheme with an equivalent fully digital case and conclude that a relatively similar performance can be achieved, however, with a drastically reduced number of RF chains - 4 ~ 8 times less (i.e., massive savings in cost and power consumption)., QC 20151217

Using a Bayesian approach, we consider the problem of recovering sparse signals under additive sparse and dense noise. Typically, sparse noise models outliers, impulse bursts or data loss. To handle sparse noise, existing methods simultaneously estimate sparse noise and sparse signal of interest. For estimating the sparse signal, without estimating the sparse noise, we construct a Relevance Vector Machine (RVM). In the RVM, sparse noise and ever present dense noise are treated through a combined noise model. Through simulations, we show the efficiency of new RVM for three applications: kernel regression, housing price prediction and compressed sensing., QC 20150512